Protandry in Pacific salmon

Transcription

1 1252 Protandry in Pacific salmon Yolanda Morbey Abstract: Protandry, the earlier arrival of males to the spawning grounds than females, has been reported in several studies of Pacific salmon (Oncorhynchus spp.). However, the reasons for protandry in salmon are poorly understood and little is known about how protandry varies among and within populations. In this study, protandry was quantified in a total of 105 years using gender-specific timing data from seven populations (one for pink salmon (O. gorbuscha), three for coho salmon (O. kisutch), two for sockeye salmon (O. nerka), and one for chinook salmon (O. tshawytscha)). Using a novel statistical procedure, protandry was found to be significant in 90% of the years and in all populations. Protandry may be part of the males strategy to maximize mating opportunities and may facilitate mate choice by females. Résumé : La protandrie, c est-à-dire l arrivée des mâles sur les frayères avant les femelles, a été signalée dans plusieurs études sur les saumons du Pacifique (Oncorhynchus spp.). Toutefois, on connaît mal les raisons de la protandrie chez les saumons, et on sait peu de choses sur les variations de ce phénomène au sein des populations et d une population à l autre. Dans notre étude, nous avons quantifié la protandrie sur un total de 105 ans à partir de données d arrivée en fonction du sexe pour sept populations (une de saumon rose, O. gorbuscha; trois de coho, O. kisutch; deux de saumon rouge, O. nerka; une de quinnat, O. tshawytscha). La nouvelle procédure statistique employée a révélé que la protandrie était significative pour 90% des années et dans toutes les populations. Il peut s agir là d un élément de la stratégie des mâles pour maximiser les possibilités d accouplement, ce qui peut faciliter le choix d un partenaire pour les femelles. [Traduit par la Rédaction] Morbey 1257 Introduction In the literature on Pacific salmon (Oncorhynchus spp.), there are frequent accounts of males migrating or arriving at the spawning grounds earlier on average than females (e.g., Pritchard 1937; Groot and Margolis 1991 and references therein; Quinn and Unwin 1993). The phenomenon of males entering a reproductive life history stage earlier than females is called protandrous reproduction, or more simply, protandry (proto-: first, andr-: males). The relative timing of males and females is an important component of any mating system, and the occurrence of protandry can provide insight into the selection pressures faced by males and females during the breeding season. The empirical and theoretical study of protandry began with and has remained focussed on insects (e.g., Wiklund and Fagerström 1977; Botterweg 1982; Iwasa et al. 1983). These authors, among others, formalized the hypothesis that protandry allows males to maximize their opportunities for mating. The first protandry models made the following simplifying assumptions about the insect mating system: (i) emergence into the reproductive phase is irreversible, (ii) males search for and mate with females during the entire emergence period, (iii) males mate several times, (iv) females mate immediately after emergence and mate only Received March 16, Accepted March 10, J15064 Y. Morbey. Behavioural Ecology Research Group, Department of Biological Sciences, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6, Canada. ymorbey@sfu.ca Can. J. Fish. Aquat. Sci. 57: (2000) once, and (v) all females mate. The prediction of protandry is robust, even when some of these conditions are relaxed (e.g., Wiklund and Fagerström 1977; Botterweg 1982; Zonneveld and Metz 1991). In a scramble competition situation where males reproduce for a longer period of time than females, initiating breeding early can guarantee males access to a greater number of mates over their lifetime than by breeding late. Pacific salmon share many features of their mating system with the above conditions (cf. Quinn et al. 1996). First, salmon are semelparous and so entry into the reproductive phase is irreversible. As a probable outcome of their semelparous life history strategy, males engage in intense intraspecific competition for access to spawning females (e.g., Foote 1990; Fleming and Gross 1994; Quinn et al. 1996). Males may spawn with several females during their breeding life span, whereas females may complete spawning over a few days followed by a longer period of nest defence (e.g., Schroder 1981; McPhee and Quinn 1998). Finally, owing to the polygynous breeding system and localized spawning areas, all females likely participate in spawning. Given that the requirements for protandry seem to be met in salmon, protandrous arrival to the spawning grounds (and initiation of spawning activities) was expected to be a general feature of salmon life histories. Although protandry is known to exist in salmon, the magnitude and frequency of protandry have not been documented. The primary objective of this study was to report on the frequency and magnitude of protandry among and within salmon populations, using a novel analytical approach. A second objective was to evaluate several competing hypotheses for protandry that have been proposed in the literature.

2 Morbey 1253 Materials and methods Data on gender-specific arrival timing to the spawning grounds were sought. Recording gender-specific arrival timing is not a management priority and usually is a byproduct of the spawner enumeration procedure used by managers. Both absolute counts (through fences or by dipnetting) and the marking component of mark recapture studies (e.g. by dipnetting or seine netting) provided gender-specific timing information. Data were excluded where samples or counts were done on the spawning grounds because these data do not bear on arrival timing per se. To examine within-population variability in protandry, multiple years of data collected in a consistent manner were required. Arrival timing data were obtained for seven populations (Table 1) that met these criteria. Details about the populations and data collection methodology can be found in Schubert and Fanos (1997) and Schubert and Tadey (1997) for the two sockeye salmon (Oncorhynchus nerka) populations, in Seiler et al. (1981) for the three coho salmon (Oncorhynchus kisutch) populations, in Fukushima and Smoker (1997) for the Auke Creek pink salmon (Oncorhynchus gorbuscha) population, and in Unwin (1986) and Unwin and Glova (1997) for the Glenariffe Stream chinook salmon (Oncorhynchus tshawytscha) population. For all but the Auke Creek data, males were classified into two age-classes (younger jacks and older adults). In addition to the timing data, annual estimates of population size, gender ratio (proportion of the population that were males), jack ratio (proportion of males that were jacks), season length, and freshwater migration distance were acquired or calculated from the available data. For the sockeye populations, the annual population sizes, gender ratios, and jack ratios were based on Schubert and Fanos (1997) and Schubert and Tadey s (1997) estimates. The cumulative counts of males and females were used to calculate annual population sizes, gender ratios, and jack ratios in the other populations. For each year in each population, season length is the last arrival date minus the first arrival date plus 1. Freshwater migration distance for each population was estimated from maps. For each year, a measure of protandry was generated. The difference in mean arrival dates of males and females was not used as a measure of protandry because the arrival distributions were not normally distributed, but rather were typically episodic. Another potential measure of protandry is the slope of the relationship between the daily gender ratio of arriving salmon and arrival date, but this measure is also inadequate. Protandry is underestimated when a few fish (with an equal gender ratio) arrive over a prolonged period late in the year and overestimated when a few fish (with excess males) arrive over a prolonged period early in the year. Instead, protandry is measured as the area (in units of days) between the cumulative percent distribution of male arrival dates (M(t)) and female arrival dates (F(t)): d Protandry ( Mt () Ft ()) t 100% where dt = 1 day. Protandry represents the number of days lag between the arrival of an equal proportion of males and females, with positive values representing the earlier arrival of males. On days when sampling did not occur, the fish counted on the next sampling day were assumed to have arrived during the previous nonsampling days according to a uniform distribution. Protandry is not standardized for season length because protandry should be a measure that is relative to the longevity of salmon on the spawning grounds and there is no evidence that longevity depends on season length. The statistical significance of the observed amount of protandry was determined using a statistical procedure based on randomization (Manly 1997). A null distribution of protandry was generated based on the empirical distribution of arrival dates. In each of 200 iterations, each fish s gender was randomized (making sure to maintain the original population gender ratio) and protandry was calculated. The observed protandry was deemed significant if it lay within the upper fifth quantile of the null distribution (i.e., a onetailed test was used). In an attempt to explain the observed variation in protandry, analysis of covariance (ANCOVA) was used to determine the effects of population size, gender ratio, season length, and population on protandry. Model selection began with a fully saturated statistical model that included each main effect and all interactions with population. Nonsignificant interaction terms were sequentially dropped from the model. If a significant interaction effect remained, the effects of population size, gender ratio, and season length on protandry were analysed separately for each population using stepwise linear regression. Jack ratio was not included in this analysis because no jacks were present in the Auke Creek pink salmon population. ANCOVA was used to test if jack ratio affected protandry differently among all populations but Auke Creek. Protandry also was compared among the seven populations using analysis of variance (ANOVA), and a post hoc means comparison test determined how the populations differed (Ryan Einot Gabriel Welsch multiple F test; SAS Institute, Inc. 1990). Using the same technique for comparing male and female timing, analyses were undertaken to determine if jacks generally arrived earlier or later than adults. SAS statistical software was used for all analyses (SAS Institute, Inc. 1990). The significance levels used for the stepwise regression models were = 0.5 for entry and = 0.05 for removal (Myers 1990). The significance level used for main effects was = 0.05 and for interaction effects was = The F statistics based on partial (type III) sum of squares for ANOVAs and t statistics (two-tailed) for the linear regression models are presented. Results Protandry was significant in most years at Auke Creek (96%, 25 of 26 years), Big Beef Creek (91%, 10 of 11 years), Birkenhead River (64%, nine of 14 years), Chilko Lake (100%, eight of eight years), Deschutes River (95%, 20 of 21 years), Sunset Falls (89%, eight of nine years), and Glenariffe Stream (88%, 14 of 16 years). In all, 90% of the years showed significant protandry (94 of 105 years). By chance alone, protandry should be significant in only five years (5% of 105 years). Protogyny was significant in one year at Sunset Falls and one year at Deschutes River. Season length affected protandry differently depending on the population (p = 0.10), so separate regression analyses were required for each population. Despite this interaction, none of the within-population analyses showed a significant effect of season length on protandry (ps > 0.14). Instead, the within-population analyses revealed a significant effect of gender ratio on protandry in Auke Creek (t 1,24 = 2.56, p = 0.02, R 2 = 0.22) (Fig. 1) but in no other population (ps > 0.23). Population size did not affect protandry in any of the seven populations (ps > 0.28). Jack ratio did not affect protandry (p > 0.5), nor did the effect of jack ratio on protandry differ among populations (p > 0.5). The magnitude of protandry differed among populations (F 6,98 = 7.90, p = ), but each population was not distinct from every other population (Fig. 2). There was no consistent pattern in the arrival schedules of jacks and adult males. Jacks arrived significantly earlier in nine of 11 years at Big Beef Creek, four of 14 years at Birkenhead River, two of eight years at Chilko River, one of 21

3 1254 Can. J. Fish. Aquat. Sci. Vol. 57, 2000 Table 1. Sources of data for the analysis of protandry and average population characteristics (±SD). Average annual population size Average annual jack ratio Average annual gender ratio (male:female) Arrival period (days) Approximate distance from ocean (km) b Total years Source a Sample or absolute count Years Population Species 26 NMFS c 0 64± ± ±8020 Auke Creek, Alaska Pink Absolute , Birkenhead River, B.C. Sockeye Sample DFO ±8 0.51± ± ± Chilko Lake, B.C. Sockeye Sample DFO ±8 0.41± ± ± WDFW 0 94± ± ± ±761 Big Beef Creek, Wash. Coho Absolute , Deschutes River, Wash. Coho Absolute WDFW 2 99± ± ± ±3045 Sunset Falls, Wash. Coho Absolute WDFW ±8 0.54± ± ±7845 Glenariffe Stream, New Zealand Chinook Absolute NIWA 100 d 156± ± ± ±893 a DFO, Department of Fisheries and Oceans, Canada; NIWA, National Institute of Water and Atmospheric Research, New Zealand; NMFS, National Marine Fisheries Service, Alaska; WDFW, Washington Department of Fisheries and Wildlife, Washington. b Distances were estimated from maps. c S.G. Taylor, NMFS, Auke Bay Laboratory, Glacier Highway, Juneau, Alaska , U.S.A., unpublished data on Auke Creek pink salmon, d Unwin (1986). years at Deschutes River, three of nine years at Sunset Falls, and seven of 16 years at Glenariffe Stream. Adult males arrived significantly later than jacks in one of 11 years at Big Beef Creek, six of 14 years at Birkenhead River, four of eight years at Chilko River, 13 of 21 years at Deschutes River, two of nine years at Sunset Falls, and three of 16 years at Glenariffe Stream. Discussion Protandrous arrival to the spawning grounds appears to be a general phenomenon in salmon. In the seven salmon populations considered, males almost always arrived at the spawning grounds earlier on average than females. Protandry occurred regardless of population attributes such as population size, jack ratio, season length, and migration distance. However, there was a tendency for protandry to increase when the gender ratio was more male biased in Auke Creek pink salmon. The magnitude of protandry may seem small (population averages range from 1 to 5 days), but nonetheless, the ubiquity of protandry requires an explanation. Our results are consistent with the mate-opportunity hypothesis, but only because the mating system of salmon is similar to the type of mating system that theoretically selects for protandry. One critical assumption must be made about the data to support the mate-opportunity hypothesis. The mate-opportunity hypothesis strictly relates to the initiation of reproductive activity, so arrival timing must correspond to spawning timing. However, breeding timing may be decoupled from arrival timing through an intervening waiting stage, and gender-specific timing adjustments may occur during this stage. For example, early-arriving chinook entering Glenariffe Stream took longer to mature (produced releasable gametes) than late-arriving chinook, and females matured more slowly than males (T.P. Quinn et al., University of Washington, unpublished data). The consequences of these processes are unclear: protandry (as measured in this study) would underestimate true protandry if female commenced spawning more slowly than males but would overestimate true protandry if both males and females matured more quickly later in the season. Unfortunately, the extent and patterns of postarrival waiting are poorly understood and could not be assessed for the other populations considered here. Under the mate-opportunity hypothesis, population density or gender ratio could influence the fitness-maximizing degree of protandry in salmon. Conditions that promote greater energy expenditure may accelerate reproductive senescence and, consequently, favour less protandry. This prediction follows logically from models that predict less protandry when mortality rates during reproduction are higher (Botterweg 1982; Iwasa et al. 1983; Parker and Courtney 1983). When gender ratios on the spawning grounds are male biased, males compete more aggressively (Schroder 1981; Fleming and Gross 1989, 1994) and presumably use energy reserves at a higher rate. Males may also compete more when the population density is greater because of closer proximity to neighbouring spawning associations (Hendry et al. 1995). Males do live for slightly fewer days when breeding densities are high (van den Berghe and Gross 1986; Hendry 1998). Estimates of population density were not available to test this prediction, but a

4 Morbey 1255 Fig. 1. Effect of gender ratio on protandry for Auke Creek pink salmon (a) shown alone and (b) in relation to all populations (solid circles, Auke Creek data; open circles, other data). For Auke Creek pink salmon, the linear equation relating protandry and gender ratio is y = x. The effect of gender ratio on protandry was not significant in the other six populations. Fig. 2. Observed interannual variation in protandry among the seven populations. Means are shown with standard deviations and sample sizes. The vertical lines at the right indicate populations with statistically similar protandry based on the post hoc means comparison test. surrogate measure, population size, did not affect protandry. There was also no indication of less protandry with a more male-biased gender ratio. An opposite result was obtained for Auke Creek pink salmon: greater protandry was observed when the gender ratio was more male biased. The mate-opportunity hypothesis could explain this pattern if a prior-resident advantage is assumed. Males that are familiar with a particular spawning locale have an advantage over an intruder when competing for a female (Foote 1990; Chebanov 1997). If females are not limiting, prior residency would not confer any advantage. If gender ratios are male biased, however, prior residency may confer an advantage during competition for limited females. The generality of this explanation must remain speculative, however, because gender ratio affected protandry in only one of the seven populations. A link between prior residency and protandry was first noted by Foote (1990), who suggested that the prior-residency advantage may explain the occurrence of protandry. However, this hypothesis is unlikely a strict alternative to the mateopportunity hypothesis. The advantage of prior residency presumably is an increase in a male s mating opportunities through competitive exclusion, but his mating opportunities also must depend on the availability of females. A different explanation for protandry is that arrival timing is constrained in some way. Perhaps males migrate upstream faster than females so that they arrive at the spawning grounds first. This would assume a gender difference in upstream migration rate, for which there is no evidence. Also, populations with a longer, and possibly more arduous, migration did not show more protandry in this study (compare Table 1 and Fig. 2). A second constraint is differential fishing pressure (i.e., more females are caught early in the run than late), but no evidence could be found to support this explanation. If factors influence the arrival timing of males independently of females (i.e., there is differential selection among the genders), protandry may be an incidental byproduct (Wiklund and Solbreck 1982). Under this hypothesis, different scenarios have been proposed for other taxa showing protandry, but none for salmon. In insects, selection for large size may be greater in females than in males, and therefore, females should terminate growth and emerge later than males (Wiklund and Solbreck 1982). Analagously, female salmon may increase their fecundity by delaying migration and continuing to feed at sea, and male salmon, if they delay migration, may suffer increased predation risk at sea that is not offset by the benefits of continued growth. However, it is questionable whether a few days of extra growth could contribute signficantly to fecundity in the much longer-lived salmon. In birds that migrate to their breeding grounds, male male competition for territories may select for the earlier arrival of males (Myers 1981). Environmental conditions would determine the arrival timing of females, independently of males. This hypothesis cannot explain protandry in salmon because males do not establish territories. Another idea, also developed for birds, is that the larger gender (usually male)

5 1256 Can. J. Fish. Aquat. Sci. Vol. 57, 2000 can tolerate harsh enviromental conditions that prevail earlier in the year, whereas the smaller gender cannot (Ketterson and Nolan 1983). However, in the current study, size did not appear to influence arrival timing because smaller jack males did not consistently arrive later than larger adult males. A different set of hypotheses assume direct selection on protandry, as does the mate-opportunity hypothesis, but from the female perspective. In insects, protandry may minimize prereproductive mortality of females while they wait to be encountered by males (Fagerström and Wiklund 1982). This is an unlikely explanation for protandry in salmon because males and females rarely would have difficulty encountering one another. Males and females migrate and arrive on the spawning grounds together, and the linear structure of the spawning grounds probably facilitates encounters. Protandry also may be a female strategy to avoid waiting costs while males prepare to breed (cf. Michener 1984). For example, dominance hierarchies may have to form before breeding is possible. In salmon, the time required to establish dominance hierarchies is probably insufficient to account for the observed degree of protandry. Rankings can be established quickly, and spawning can occur within hours (personal observation). However, protandry also may benefit females by facilitating mate choice while minimizing waiting costs (cf. Wang et al. 1990). If females delay arrival, increased male male competition during the formation of dominance hierarchies may single out the higher-quality males. This hypothesis may have merit because female salmon do exhibit preference for males based on quality measures such as size (Foote 1989). Based on the natural history of salmon, sexual selection likely acts directly on protandry. From the male s perspective, protandry is consistent with a strategy to maximize opportunities for mating. From the female s perspective, protandry may facilitate mate choice. However, further study is required to conclusively reject the competing hypotheses and to understand why protandry varies among and within populations. For example, it is unclear why some populations, such as Big Beef Creek coho salmon and Birkenhead River sockeye salmon, differ in protandry. Perhaps population differences in species, breeding longevity, arrival-timing constraints, or postarrival waiting are responsible. Another unexplained pattern is the effect of gender ratio on protandry in Auke Creek pink salmon. In addition to understanding the adaptive significance of protandry, the mechanisms producing protandry need to be understood. The hypotheses explain why protandry exists but not how it comes about (i.e., the proximate mechanisms). If timing cannot be adjusted to the ecological and social factors that define optimal protandry, reproductive output could be compromised. Acknowledgments I am especially grateful to Jerry Taylor of the National Marine Fisheries Service, Alaska (NMFS), Neil Schubert and Tracy Cone of Canada s Department of Fisheries and Oceans (DFO), Dave Seiler and Steve Neuhauser of the Washington State Department of Fisheries and Wildlife (WDFW), and Martin Unwin of the National Institute of Water and Atmospheric Research, New Zealand (NIWA), for providing the data that I used in the analyses. This project would have been impossible without their help and the cooperation of NMFS, DFO, WDFW, and NIWA. Bernie Crespi, Larry Dill, Andrew Hendry, Martin Unwin, Ron Ydenberg, and two anonymous reviewers all provided excellent advice on how to improve the manuscript. Ian Hamilton, Stephanie Hazlitt, Andrew Hendry, Don Hugie, and Sharilynn Wardrop contributed helpful advice about measuring protandry, and Martin Unwin suggested the measure that I present here. This research forms part of the Ph.D. dissertation of Y.M. Funding to Y.M. was provided by Natural Sciences and Engineering Research Council of Canada and Simon Fraser University graduate fellowships. References Botterweg, P.F Protandry in the pine looper, Bupalus piniarius (Lep., Geometridae); an explanatory model. Neth. J. Zool. 32: Chebanov, N.A Role of the prior resident effect in formation of the dominance subordination relations and determination of the reproductive success values in Pacific salmons. J. Ichthyol. 37: Fagerström, T., and Wiklund, C Why do males emerge before females? Protandry as a mating strategy in male and female butterflies. Oecologia, 52: Fleming, I.A., and Gross, M.R Evolution of adult female life history and morphology in a Pacific salmon (coho: Oncorhynchus kisutch). Evolution, 43: Fleming, I.A., and Gross, M.R Breeding competition in a Pacific salmon (coho: Oncorhynchus kisutch): measures of natural and sexual selection. Evolution, 48: Foote, C.J Female mate preference in Pacific salmon. Anim. Behav. 38: Foote, C.J An experimental comparison of male and female spawning territoriality in a Pacific salmon. Behaviour, 115: Fukushima, M., and Smoker, W.W Determinants of stream life, spawning efficiency, and spawning habitat in pink salmon in the Auke Lake system, Alaska. Can. J. Fish. Aquat. Sci. 54: Groot, C., and Margolis, L. (Editors) Pacific salmon life histories. University of British Columbia Press, Vancouver, B.C. Hendry, A.P Reproductive energetics of Pacific salmon: strategies, tactics, and trade-offs. Ph.D. thesis, University of Washington, Seattle, Wash. Hendry, A.P., Leonetti, F.E., and Quinn, T.P Spatial and temporal isolating mechanisms: the formation of discrete breeding aggregations of sockeye salmon (Oncorhynchus nerka). Can. J. Zool. 73: Iwasa, Y., Odendaal, F.J., Murphy, D.D., Ehrlich, P.R., and Launer, A.E Emergence patterns in male butterflies: a hypothesis and a test. Theor. Popul. Biol. 23: Ketterson, E.D., and Nolan, V., Jr The evolution of differential bird migration. In Current ornithology. Vol. 1. Edited by R.F. Johnston. Plenum Press, New York. pp Manly, B.F.J Randomization, bootstrap and Monte Carlo methods in biology. 2nd ed. Chapman and Hall, New York. McPhee, M.V., and Quinn, T.P Factors affecting the duration of nest defense and reproductive lifespan of female sockeye salmon, Oncorhynchus nerka. Environ. Biol. Fishes, 51: Michener, G.R Spring emergence schedules and vernal behavior of Richardson s ground squirrels: why do males emerge from hibernation before females? Behav. Ecol. Sociobiol. 14:

6 Morbey 1257 Myers, J.P A test of three hypotheses for latitudinal segregation of the sexes in wintering birds. Can. J. Zool. 59: Myers, R.H Classical and modern regression with applications. 2nd ed. PWS-KENT, Boston, Mass. Parker, G.A., and Courtney, S.P Seasonal incidence: adaptive variation in the timing of life history stages. J. Theor. Biol. 105: Pritchard, A.L Variation in the time of run, sex proportions, size and egg content of adult pink salmon (Oncorhynchus gorbuscha) at McClinton Creek, Masset Inlet, B.C. J. Biol. Board Can. 3: Quinn, T.P., and Unwin, M.J Variation in life history patterns among New Zealand chinook salmon (Oncorhynchus tshawytscha) populations. Can. J. Fish. Aquat. Sci. 50: Quinn, T.P., Adkison, M.D., and Ward, M.B Behavioral tactics of male sockeye salmon (Oncorhynchus nerka) under varying operational sex ratios. Ethology, 102: SAS Institute, Inc SAS/STAT user s guide. SAS Institute, Inc., Cary, N.C. Schroder, S.L The role of sexual selection in determining overall mating patterns and mate choice in chum salmon. Ph.D. thesis, University of Washington, Seattle, Wash. Schubert, N.D., and Fanos, B.P Estimation of the 1994 Chilko River and Chilko Lake system sockeye salmon (Oncorhynchus nerka) escapement. Can. Manuscr. Rep. Fish. Aquat. Sci. No Schubert, N.D., and Tadey, J.A Estimation of the 1994 Birkenhead River sockeye salmon (Oncorhynchus nerka) escapement. Can. Manuscr. Rep. Fish. Aquat. Sci. No Seiler, D., Neuhauser, S., and Ackley, M Upstream/downstream salmonid trapping project, State Wash. Dep. Fish. Prog. Rep. No Unwin, M.J Stream residence time, size characteristics, and migration patterns of juvenile chinook salmon (Oncorhynchus tshawytscha) from a tributary of the Rakaia River, New Zealand. N.Z. J. Mar. Freshwater Res. 20: Unwin, M.J., and Glova, G.J Changes in life history parameters in a naturally spawning population of chinook salmon (Oncorhynchus tshawytscha) associated with releases of hatchery-reared fish. Can. J. Fish. Aquat. Sci. 54: van den Berghe, E.P., and Gross, M.R Length of breeding life of coho salmon (Oncorhynchus kisutch). Can. J. Zool. 64: Wang, G-Y., Greenfield, M.D., and Shelly, T.E Inter-male competition for high-quality host-plants: the evolution of protandry in a territorial grasshopper. Behav. Ecol. Sociobiol. 27: Wiklund, C., and Fagerström, T Why do males emerge before females? Oecologia, 31: Wiklund, C., and Solbreck, C Adaptive versus incidental explanations for the occurrence of protandry in a butterfly, Leptidea sinapis L. Evolution, 36: Zonneveld, C., and J.A.J. Metz Models on butterfly protandry: virgin females are at risk to die. Theor. Popul. Biol. 40: